CN104321020A - Tumor classification based on an analysis of a related ultrasonic attenuation map - Google Patents

Tumor classification based on an analysis of a related ultrasonic attenuation map Download PDF

Info

Publication number
CN104321020A
CN104321020A CN201380027801.0A CN201380027801A CN104321020A CN 104321020 A CN104321020 A CN 104321020A CN 201380027801 A CN201380027801 A CN 201380027801A CN 104321020 A CN104321020 A CN 104321020A
Authority
CN
China
Prior art keywords
described
roi
suspicious
image
decay
Prior art date
Application number
CN201380027801.0A
Other languages
Chinese (zh)
Other versions
CN104321020B (en
Inventor
S·格莱希曼
E·瓦拉施
Original Assignee
国际商业机器公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to US13/558,372 priority Critical
Priority to US13/558,372 priority patent/US8781202B2/en
Application filed by 国际商业机器公司 filed Critical 国际商业机器公司
Priority to PCT/CA2013/050486 priority patent/WO2014015424A1/en
Publication of CN104321020A publication Critical patent/CN104321020A/en
Application granted granted Critical
Publication of CN104321020B publication Critical patent/CN104321020B/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0012Biomedical image inspection
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10132Ultrasound image
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30004Biomedical image processing
    • G06T2207/30068Mammography; Breast

Abstract

A computerized method of classifying at least one suspicious region of interest (ROI) in an ultrasonic attenuation image mapping tissue of a patient. The method comprises receiving an US image of an tissue, identifying a suspicious region of interest (ROI) in the US image, generating an attenuation map of the suspicious ROI, measuring, according to an analysis of the attenuation map, at least one attenuation feature of at least one of the suspicious ROI and at least one sub region in the suspicious ROI, and classifying the suspicious ROI according to the at least one attenuation feature.

Description

Based on the staging of the analysis of associated ultrasonic decay pattern

Technical field

The present invention relates to computer-aided diagnosis in its some embodiments, and more specifically but its ground non-relates to the system and method for the suspicious region in the ultrasonoscopy of homogeneity or the non-homogeneous tissue assessing such as breast tissue and so on.

Background technology

The ultrasonic cancer diagnosis being widely used in breast carcinoma.Worldwide, breast carcinoma comprises nearly 30% of all cancer diagnosis in women.Mammography is the most common current therapy for screening and detect breast carcinoma; But the most breast lesion found in mammography is only optimum.In order to improve specificity, doctor often uses ultrasonic (US) imaging inspection suspicious lesions.But even if when using both mammography and US, the biopsy of about 80% is still optimum.

In recent years, there is kinds cancer diagnostic tool, see Nover, A.B., Jagtap, S., Anjum, W., Yegingil, H., Shih, W., Shih, W. and Brooks, A.D. the Modern Breast Cancer Detection:A Technological Review delivered in the international weekly (2009) of biomedical imaging, and Sehgal, C.M., Weinstein, S.P., Arger, and the A Review of Breast Ultrasound that delivers in biological and neoplasia 11 (2) (2006) 113-123 at mammary gland of Conant, E.F. P.H..

Certain cancers diagnostic tool comprises area of computer aided (CAD) system.These systems calculate various breast image feature usually to distinguish pernicious and benign tumor.Breast image feature comprises the shape (be estimated as describe possible tumor) of suspicious region in image, the texture of suspicious region and the acoustic properties of suspicious region.

Summary of the invention

According to some embodiments of the present invention, provide the Computerized method that at least one suspicious regions of interest (ROI) in a kind of ultrasonic attenuation image of the tissue to mapping patient is classified.The method comprises suspicious regions of interest (ROI) in the US image of tissue receiving, mark US image, generates the decay pattern of suspicious ROI, according to measuring at least one decay characteristics of at least one the suspicious ROI in suspicious ROI and at least one subregion in suspicious ROI to the analysis of decay pattern and classifying to suspicious ROI according at least one decay characteristics.

According to some embodiments of the present invention, provide the system that in a kind of ultrasonic attenuation image of the tissue to mapping patient, at least one suspicious regions of interest (ROI) is classified.System comprises the ROI module of suspicious regions of interest (ROI) in processor, the input module of US image of tissue receiving, mark US image, the attenuation map module generating the decay pattern of ROI and the sort module of classifying to suspicious ROI according at least one decay characteristics of at least one subregion in suspicious ROI, identifies at least one decay characteristics according to the analysis of decay pattern.

According to some embodiments of the present invention, provide a kind of computer program of classifying at least one suspicious regions of interest (ROI) in the ultrasonic attenuation image of the tissue to mapping patient.Computer program comprises non-transient computer-readable recording medium, for the first programmed instruction of the US image of tissue receiving, for identifying the second programmed instruction of suspicious regions of interest (ROI) in US image, for generating the 3rd programmed instruction of the decay pattern of suspicious ROI, for the 4th programmed instruction according at least one decay characteristics of at least one subregion at least one and the suspicious ROI of the suspicious ROI of the analysis to measure of decay pattern, and according to the 5th programmed instruction that at least one decay characteristics is classified to suspicious ROI.Non-transient computer-readable recording medium stores first, second, third, fourth and fifth programmed instruction.

Unless otherwise defined, all technology used herein and/or scientific terminology have the identical meanings as those skilled in the art understand usually.Similar or equivalent method described herein and goods and materials can be used in the practice or test of embodiments of the invention, exemplary method and/or goods and materials are described below.In the event of a conflict, patent specification (comprising restriction) will control.In addition, goods and materials, method and example are only illustrative, and not intended to be must limit.

Accompanying drawing explanation

Patent or application documents comprise at least one color accompanying drawing.Patent Office provide based on request and the payment of necessary expenses there is color accompanying drawing this patent or patent application disclosed in copy.

By means of only accompanying drawing appended by exemplary reference, some embodiments of the present invention are described at this.Now specifically see accompanying drawing, should emphasize, shown particular content is only example and is the object discussed for the signal of embodiments of the invention.Thus, how the description of reference accompanying drawing can put into practice embodiments of the invention is for those skilled in the art obvious.

In the accompanying drawings:

Fig. 1 is the flow chart of classifying to the one or more suspicious ROI mapped in the homogeneity of patient or the US image of non-homogeneous tissue based on the analysis of ultrasonic attenuation feature;

Fig. 2 be according to some embodiments of the present invention such as by implementing the method described in Fig. 1 schematic diagram to the parts of the system that ROI in the ultrasonoscopy mapping homogeneity or non-homogeneous tissue classifies;

Fig. 3 A-Fig. 3 C be according to certain pixel of some embodiments of the present invention near example segment, be applied in figure speckle (mask) on example segment and removing the figure of treated figure speckle of peripheral picture group (blob);

Fig. 4 A-Fig. 4 C is the US image according to some embodiments of the present invention, the US image with the suspicious ROI of hand labeled and the decay pattern generated according to the suspicious ROI described in Fig. 4 B;

Fig. 5 A-Fig. 5 D is some examples of the decay pattern describing carcinoid ROI;

Fig. 5 E-Fig. 5 H is some examples of the decay pattern of the ROI describing malignant tumor;

Fig. 6 A-Fig. 6 D is description benign tumor according to some embodiments of the present invention and the US image graph district corresponding with suspicious ROI, have the smoothed version of the decay pattern indicating the decay pattern of the wire tag of tumor boundaries, the smoothed version of decay pattern and have the color region representing Uniform attenuation region; And

Fig. 7 A-Fig. 7 E describes the software code according to the analysis for realizing decay characteristics of some embodiments.

Detailed description of the invention

The present invention relates to computer-aided diagnosis (CAD) in its some embodiments, and more specifically but non-exclusively relate to the method and system of the classification of the suspicious region in the ultrasonoscopy of tissue.

According to some embodiments of the present invention, be provided for the decay pattern by generating ROI and analyze the one or more decay characteristics method and system of classifying to homogeneity or non-homogeneous tissue extracted from decay pattern, this homogeneity or non-homogeneous organize the breast tissue of such as description in the suspicious regions of interest (ROI) of ultrasonic (US) image of such as B-scan image and so on, parathyroid tissue, colon and prostata tissue.Can be able to be the ROI of two dimension or volume in US image by user's labelling and/or Automatic Logos.

Alternatively, the feature analyzed relates to the homogeneity of yardstick, pad value and pad value in the subregion in ROI.The current potential pattern of the optimum and/or malignant tumor in the tissue image in these features instruction ROI.

Alternatively, when considering neighborhood pixels value, carry out generation of attenuation map by calculating decay pixel value alternatively.

Before illustrating at least one embodiment of the present invention, should be appreciated that, the invention is not restricted to it and hereafter describing and/or the parts shown in accompanying drawing and/or example and/or the structure of method and the details of layout.The present invention can perform or put into practice some other embodiment in every way.

As will be understood by the skilled person, aspects more of the present invention can be presented as system, method or computer program.Correspondingly, aspects more of the present invention can take to be entirely hardware embodiment, the form of the embodiment of be entirely software implementation (comprising firmware, resident software, microcode etc.) or integration software and some aspects of hardware, and this hardware aspect is totally called as " circuit ", " module " or " system " in this article.In addition, the form of the computer program embodied in one or more computer-readable medium can be taked in aspects more of the present invention, and this computer-readable medium has the computer readable program code embodied thereon.

Any combination of one or more computer-readable medium can be utilized.Computer-readable medium can be computer-readable signal media or computer-readable recording medium.Computer-readable recording medium can such as but not limited to any appropriate combination of ground electronics, magnetic, optical, electrical magnetic, infrared or semiconductor system, device or equipment or aforementioned item.The how concrete example (non-exhaustive list) of computer-readable recording medium can comprise following item: the random suitable combination with the electrical connection of one or more wiring, portable computer video disc, hard disk, random access memory (RAM), read only memory (ROM), Erasable Programmable Read Only Memory EPROM (EPROM or flash memory), optical fiber, Portable, compact dish read only memory (CD-ROM), optical storage apparatus, magnetic storage apparatus or aforementioned item.In situation herein, computer-readable recording medium can any tangible medium, and it can comprise or store the program for instruction execution system, device or equipment.

Computer-readable signal media can comprise the data signal of propagation, and it has the computer readable program code embodied wherein, such as, in a base band or as the part of carrier wave.The signal of this propagation can adopt various forms, such as but not limited to, electromagnetism, light or its random suitable combination.Computer-readable signal media can be any computer-readable medium, and it is not computer-readable recording medium and and instruction executive system, device or equipment can combines for communication, propagates or transmission procedure.

Any suitable medium can be used to transmit the program code embodied on a computer-readable medium, and this medium includes but not limited to any appropriate combination of wireless, wired, optical fiber cable, RF etc. or aforementioned item.

Computer program code for performing aspects more of the present invention can be write with any combination of one or more programming languages, and this programming language comprises in the face of Object Programme language (such as Java, Smalltalk, C++ etc.) and conventional process programming language (such as " C " programming language or similar programming language).Program code can completely on the user computer, part on the user computer as freestanding software bag, part on the user computer and part perform on remote computer or server on the remote computer or completely.In the situation of the latter, remote computer can be connected to the computer of user by the network of any type, this any type network comprises LAN (LAN) or wide area network (WAN), or connection can connect externally computer (such as by using the Internet of ISP).

Below with reference to the illustrated flow chart of the method according to some embodiments of the present invention, device (system) and computer program and/or block diagram, aspects more of the present invention are described.Be appreciated that the combination that can be performed the frame in illustrated flow process and/or block diagram and flow chart and/or block diagram by computer program instructions.These computer program instructions can be supplied to the processor of general purpose computer, special-purpose computer or other programmable data blood processor to produce machine, make call instruction generation device for performing the function/action of specifying in flow chart and/or block diagram when the processor via computer or other programmable data blood processor performs.

Also these computer program instructions can be stored in computer-readable medium, it can order computer, other programmable data blood processor or miscellaneous equipment work in a specific way, make to be stored in instruction in computer-readable medium and produce the goods comprising instruction, it implements the function/action of specifying in flow process and/or block diagram.

Computer program instructions can also to be loaded on computer, other programmable data blood processor or miscellaneous equipment with the step that the operating procedure making to perform on computer, other programmable data blood processor or miscellaneous equipment some performs to produce computer, makes the instruction performed on computer or other programmable device be provided for the function/action realizing specifying in flow chart and/or block diagram or frame.

Referring now to Fig. 1, Fig. 1, be there is the analysis of decay characteristics in the US image of the tissue of homogeneity or non-homogeneous (being called for short non-homogeneous tissue herein) ultrasonic radiation absorption to the flow chart of the method 100 that one or more suspicious ROI classifies according to some embodiments of the present invention based on mapping.Decay characteristics is the feature (also referred to as picture group) of suspicious ROI and/or its subregion.These decay characteristics extract from the decay pattern of suspicious ROI alternatively.

Referring now to Fig. 2, Fig. 2 is that this non-homogeneous organizes such as breast tissue, colon and prostata tissue according to some embodiments of the present invention such as by realizing the method described in Fig. 1 schematic diagram to the parts of the system that the one or more ROI mapped in the ultrasonoscopy of non-homogeneous tissue classify.As shown in the figure, software part comprise for alternatively by network from ultrasound therapy 71 and/or hold storehouse 72 and receive the input interface 61 of ultrasonoscopy, ROI module 63, attenuation map module 64 and sort module 65.

Ultrasound therapy 71 is that B scans ultrasound therapy alternatively, and it generates ultrasonoscopy based on back scattering.

Method and system is based on the analysis of acoustic features, and namely acoustics tissue attenuation is measured and is used for distinguishing between optimum and malignant tumor.Ultrasonic (US) transmission of tomography mode is used to carry out computation organization's decay.

First, as illustrated at 101 places, provide ultrasonic (US) image of the non-homogeneous tissue of such as breast tissue and so on from ultrasound therapy 71 (such as B scans therapy) and/or calm storehouse 72 (such as medical data base).US image is conventional B scanogram alternatively, and it is taken when not revising the hardware of image acquisition procedures and/or use.

Now, as illustrated at 102 places, in US image, suspicious ROI is identified.Suspicious ROI can by doctor and/or operator's hand labeled.Such as, ROI mark module 3 is set to present graphical user interface (GUI), it allows user such as by drawing line nearby with the ROI presented of labelling US image on the display of client, this client is desktop computer, laptop computer, tablet PC etc. such as, the line described in Fig. 4 B on the US image that this line such as presents in figure 4 c.In another embodiment, automatically ROI is identified by the image procossing of US image, for example, see investigate about US image Zhong Tu district: Noble and Boukerroui, " Ultrasound Image Segmentation:A Survey ", IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL.25, NO.8, in August, 2006.

Now, as shown in 103, the ROI of labelling and its near zone is analyzed to produce decay pattern by attenuation map module 64 alternatively.Alternatively, the local attenuation value by calculating each pixel in ROI produces decay pattern.Local attenuation value is estimated alternatively for each pixel according to around block.Suppose that the decay of block is uniform, alternatively except the exposed area (outlier) (such as pixel is significantly different from the place of the center pixel of block) be removed.

Alternatively, such as decay pattern is presented over the display.Decay pattern allows doctor to distinguish between optimum and malignant tumor, and reduces the number of the current redundancy biopsy be performed thus.

Alternatively, decay pattern is produced in the following manner:

For each pixel, in the roi and alternatively in its vicinity:

Limit frame size LxP around;

Limit the frame size with surrounding and comprise the figure speckle of frame of the surrounding of pixel and the multiple pixel of surrounding thereof, such as, making pixel be positioned at the center of the frame of surrounding;

The frame pad value of frame is around estimated according to the pad value of the pixel in figure speckle; And

Frame pad value is assigned as the decay map values of pixel.

Alternatively, frame pad value is the frame average attenuation using method of least square to calculate.

Alternatively, in US image n-th row and m capable in pixel intensity restriction as follows:

Equation 1: E m , n = E 0 σ m , n exp ( - 2 Δ Σ k = 1 m α k , n )

Wherein, E 0expression initial amplitude, △ represent size and the σ of pixel m; nand α k,nrepresent back scattering and attenuation quotient respectively.Do not lose universality, suppose Δ=1.This equals the simple measurement unit changing decay.

At suspicious ROI with estimate α to each pixel in its near zone alternatively k,n.Decay and backscattering coefficient are assumed to be constant in the little near zone of each pixel.As mentioned above, near zone is restricted to the frame of the surrounding of LxP the pixel (such as 65x17 pixel) had around each pixel.If decay is uniform in this frame, then equation 1 is reduced to:

Equation 2:E j, i=E ie -2 (j-1) α

Wherein, E j,irepresent the intensity in (j, i) pixel of frame, and α represents the constant attenuation of the pixel of frame.

Alternatively, use figure speckle determines that pixel has the position of constant attenuation.Alternatively, for pixel (there are near such as center pixel all pixels in the frame of about 3 decibels of (dB) interior intensity) the marked graph speckle in all scopes.Then, the non-zero pixels multi dimensional object of the periphery connection being called as non-zero picture group is in this article removed from figure speckle.Such as, it is removed apart from any non-zero picture group of main picture group more than 4 pixels.

Because figure speckle indicates the subregion with homogeneity decay, the model therefore in equation 2 is suitable for all pixels indicated by figure speckle.Therefore, cost function can be defined as the summation that wherein constant attenuation supposes all pixels of the variance of rational frame.Then, equation 2 is only applied to the pixel on this figure speckle.

Fig. 3 A-Fig. 3 C is example frame around certain pixel 301, be applied in figure speckle on this example frame and at the image removing the treated figure speckle after peripheral picture group, vide infra:

Equation 3: C ( E i , α ) = Σ i = 1 P Σ j ∈ Ω i ( E j , i - E i e - 2 ( j - 1 ) α ) 2

Wherein Ω irepresent the i-th row of figure speckle.Target is the E finding to minimize this cost function iand α.Alternatively, technical scheme is limited to its first approximation.This is suitable under being limited in the hypothesis of α <<1, makes the index in equation 3 be little.Because the pad value of most of biological tissue is seldom on every pixel 0.01 napier, therefore this hypothesis is effective.The first approximation of equation 3 is:

Equation 4: C ( E i , &alpha; ) = &Sigma; i = 1 P &Sigma; j &Element; &Omega; i ( E j , i - E i ( 1 - 2 ( j - 1 ) &alpha; ) ) 2

Wherein k irepresent the number of the non-zero pixels on the i-th row of figure speckle, and define following content:

a i = 2 k i &Sigma; j &Element; &Omega; i ( j - 1 ) , b i = 1 k i &Sigma; j &Element; &Omega; i E j , i , c i = 1 k i &Sigma; j &Element; &Omega; i ( j - 1 ) E j , i - a i b i

Using this labelling, being easy to find out, when being set to 0 according to E iand use first approximation, C (E i, α) derivative produce:

Equation 5: E ^ i = b i - 2 c i &alpha;

Equation 5 is substituted into equation 4, and keeps only first rank of α, produce the simple chi square function of α, its single global minimum is in following formula:

Equation 6: &alpha; ^ = &Sigma; i = 1 P D i T B i &Sigma; i = 1 P B i T B i

Wherein Ω (j, i)represent (j, the i) pixel in figure speckle, and

D i = ( b i - E 1 , i , . . . , b i - E L , i ) T , B i = 2 ( &Omega; 1 , i c i , . . . , &Omega; L , i ( c i - ( L - 1 ) b i ) . . . ) T

The pad value estimated only be assigned to center pixel and be not assigned to whole frame.Perform this calculating for each pixel in the roi and around to produce decay pattern.Such as, Fig. 4 C is the image of the decay pattern generated according to suspicious ROI, labelling on the US image that this suspicious ROI is presented in Figure 4 A by doctor in figure 4b.

Now, as shown in 104, the one or more decay characteristics in suspicious ROI and/or the subregion in suspicious ROI can be calculated according to decay pattern.Alternatively, decay characteristics is one or more subregions of suspicious ROI.

Referring now to the example that Fig. 5 A-Fig. 5 D, Fig. 5 A-Fig. 5 D is the decay pattern describing carcinoid ROI, Fig. 5 E-Fig. 5 H is the example of the decay pattern of the ROI describing malignant tumor.Each ROI is gone out by coil.As observed in Fig. 5 E-Fig. 5 H, malignant tumor has relatively large height decay speckle, and population structure is non-homogeneous.As illustrated in these figures, due to the distortion of such as time gain compensation (TGC) and so on, therefore decay pattern is relative and nisi, and wherein, the tissue for health is expected 0 decay and expects higher pad value for malignant tumor.It should be noted that identical scheme may be used for the image obtained when not having TGC.

Also see the characteristic set that may be used for classifying to suspicious ROI:

Table 1

Alternatively, in order to quantization characteristic 1 and 2, use H-summit (H-maxima) conversion to process decay pattern, H-maxima conversion suppresses (mild) summit of appropriateness.For feature number 1, such as, identified the subregion of relatively high decay by threshold application on level and smooth figure (being chosen as fixing, (10-3 napier/pixel)), use too small area filtering picture group simultaneously.For feature number 2, analyze two regions in level and smooth figure with Uniform attenuation alternatively.First area comprises all pixels that its value equals the intermediate value of level and smooth decay pattern, and second area is defined as all pixels that its value equals the intermediate value of not pixel in the first region similarly.Fig. 5 A-Fig. 5 D is the example image in suspicious ROI in US figure and corresponding Uniform attenuation region.Fig. 6 A describes carcinoid US image graph district, and correspond to suspicious ROI, Fig. 6 B describes the decay pattern wherein describing wire tag tumor boundaries, and Fig. 6 C and Fig. 6 D is the smoothed version of decay pattern and has the smoothed version of decay pattern of the color region representing Uniform attenuation region 70.

Alternatively, according to pad value identification characteristics 3 in image.Feature 4 checks the region of wherein decaying close to maximum.As a result, filtering artifact, ignores the maximum only occurred in isolate pixels in a small amount simultaneously.

Alternatively, feature depends on the size of ROI.Such as, the most I of feature 3 accepts the parameter of intensity and H-maxima conversion is less when the less tumor of reply.This is due to the following facts: that, for less tumor, decay behavior accuracy rate is lower.

Now, as shown in 105, according to one or more feature (such as according to characteristics combination), suspicious ROI is classified.Alternatively, feature is weighted.Classification can be binary, such as, distinguish between malignant tumor and benign tumor.The directional point of classification can be tumor be optimum and/or pernicious probability.Alternatively, classification shows user, such as, as the transparent print (overlay) in the display of US image.

Alternatively, as shown in 106, repeat the process shown in 102-105 iteratively, at every turn on another suspicious ROI.In this embodiment, can multiple suspicious ROI individually in analysis chart picture.

Fig. 7 A-Fig. 7 E depicts the software code of the above-mentioned analysis for performing decay characteristics according to some embodiments of the present invention.

According to some embodiments of the present invention, US image is three-dimensional (3D) image, such as, use the imaging of 3D US therapy.In this kind of embodiment, suspicious ROI is the volume ROI with the width of mark, length and the degree of depth, its automatically, by doctor and/or another operator semi-automatic (such as soliciting the Automatic Logos that user agrees to) or manually identify.

Alternatively, as shown in 107, repeat the process described in 102-105 and optional 106 iteratively, at every turn on another sectioning image of suspicious volume ROI.In this kind of embodiment, can multiple suspicious ROI individually in analysis chart picture.In another embodiment, each section performs 103 individually, and perform figure district, signature identification and classification on the suspicious ROI of volume covering 3D tumor.

Alternatively, as shown in 108, such as, show the classification of one or more suspicious EOI over the display to user, such as, as the transparent print of information and/or the report for user's generation that use corresponding US image shows.Alternatively, perform same analysis for some views with identical tumor, wherein perform and throw choosing to combine the structure for each view.

Said method is used in IC chip manufacture.

Flow chart in accompanying drawing and block diagram show system according to various embodiments of the present invention, the framework of possible implementation of method and computer program code, functional and operation.Thus, each frame in flow chart or block diagram can the part of representation module, field or code, and it comprises the one or more executable instructions for realizing logic function.It should be noted that in some Alternate implementations, in frame, the function of labelling can perform with the order different from shown in figure.Such as, in fact two blocks in succession shown can perform substantially concurrently, or frame can perform sometimes in reverse order, and this depends on involved functional.Should also be noted that the combination of the frame in block diagram and/or the illustrated each frame of flow process and block diagram and/or flow process diagram can realize by based on the system of specialized hardware or the combination of specialized hardware and computer instruction, this system performs special function or action.

For illustrated object but and not intended to be limit or illustrate the description of various embodiment of the present invention with being limited to the disclosed embodiments.Many modifications and variations are obvious for those skilled in the art, and do not depart from the scope and spirit of described embodiment.Selection of terms used herein is the principle that embodiment is described best, compared to technology existing in market, and actual application or technological improvement or make those skilled in the art understand embodiment described herein.

Be appreciated that and be derived from the patent life period of this application, will many relevant method and systems be occurred, and the scope of term processor, module, US image, B-scan and ultrasound therapy is intended to all this kind of new technology comprising deduction.

Term " about " used herein refers to 10%.

Term " comprises ", " comprising ", " containing ", " having " and morphological change thereof mean " including but not limited to ".This term comprises term " by ... composition " and " being substantially made up of ... .. ".

Expressing " being substantially made up of ... .. " means composition or method can comprise additional composition and/or step, but does not only change in fact the fundamental sum novel feature of composition in claims or method at additional composition and/or step.

As used herein, singulative " ", " one " comprise plural form, unless the context clearly indicates otherwise.Such as, term " a kind of compound " or " at least one compound " can comprise multiple compounds, comprise its mixture.

Word " example " is in this article for meaning " as example, example or diagram ".Any embodiment being described as " example " must not be configured to compared to other embodiment to be preferred or favourable and/or to get rid of being incorporated to of feature of other embodiment.

Term " optional " is in this article for meaning " providing in certain embodiments and do not provide in further embodiments ".Any specific embodiment of the present invention can comprise multiple " optional " feature, unless the conflict of these features.

Run through the application in full, various embodiment of the present invention can be shown with range format.Should be appreciated that, the description of range format only for convenient and simple and clear, and should not be configured to the immutable restriction of scope of the present invention.Correspondingly, the description of scope should be considered to have concrete disclosed likely subrange and the independent digital value within the scope of this.Such as, the scope of description such as should be considered to have concrete disclosed subrange from 1 to 6, such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, such as, from 3 to 6 etc. and independent numeral within the scope of this, 1,2,3,4,5 and 6.Regardless of the width of scope, this can be suitable for.

No matter when indicating range in this article, all to mean to be included in the scope of instruction any quotes numerical value (mark or integer).Express that " scope between the first designation number and the second designation number and from the first designation number to the scope of the second designation number " is mutual in this article to be used and be intended to all marks and the integer numerical value that comprise the first designation number and the second designation number and period.

Of the present invention also can the combination in single embodiment in some that the object being appreciated that for clearness describes in the situation of independent embodiment provides.On the contrary, the various features described in single embodiment for simple object of the present invention also can provide or individually with suitable sub-portfolio or describe mode suitable in embodiment at any other of the present invention and provide.Some feature described in the situation of various embodiment is not considered to the basic feature of these embodiments, unless this embodiment can not operate when not having this element.

Although describe the present invention in conjunction with specific embodiments of the invention, obvious many alternative, modifications and variations are obvious for those skilled in the art.Correspondingly, the present invention be intended to comprise fall into appended claims spirit and broad scope in all this kind of alternative, modifications and variations.

The all publication, patents and patent applications mentioned in this description are incorporated herein by reference in their entirety at this, as each independent document, patent or patent application specifically and being designated as individually and being incorporated to by reference herein.In addition, quoting or identifying of any reference in the application should not be interpreted as admitting that this is referenced as prior art of the present invention.In open scope herein, they should not be interpreted as inevitable restriction.

Claims (20)

1., to the Computerized method that at least one suspicious regions of interest (ROI) in the ultrasonic attenuation image of the tissue of mapping patient is classified, comprising:
The US image of tissue receiving;
Suspicious regions of interest (ROI) is identified in described US image;
Generate the decay pattern of described suspicious ROI;
At least one decay characteristics of at least one of at least one subregion in described suspicious ROI and described suspicious ROI is measured according to the analysis of described decay pattern; And
According at least one decay characteristics described, described suspicious ROI is classified.
2. Computerized method according to claim 1, at least one decay characteristics wherein said is the ratio between the area of described suspicious ROI and multiple subregions with the pixel pad value on predeterminated level of described suspicious ROI.
3. Computerized method according to claim 1, wherein said tissue is the non-homogeneous tissue with non-homogeneous ultrasonic radiation absorption.
4. Computerized method according to claim 1, at least one decay characteristics wherein said is the yardstick comprising the part of multiple subregions with relatively uniform pixel pad value of described suspicious ROI.
5. Computerized method according to claim 1, at least one decay characteristics wherein said is the maximum attenuation value of a part of described suspicious ROI.
6. Computerized method according to claim 1, at least one decay characteristics wherein said is the yardstick comprising the part of accumulation pixel pad value on the maximum attenuation value of the pixel of described suspicious ROI of described suspicious ROI.
7. Computerized method according to claim 1, at least one decay characteristics wherein said be described suspicious ROI comprise the yardstick that burden amasss the part of pixel pad value.
8. Computerized method according to claim 1, at least one decay characteristics wherein said is the size of described suspicious ROI.
9. Computerized method according to claim 1, wherein said generation comprises each pixel for described suspicious ROI:
Limit the surrounding's frame comprising respective pixel and the multiple pixel of surrounding thereof, and
The decay map values according to the combination calculation of the pad value of described multiple pixel is assigned to described respective pixel.
10. Computerized method according to claim 1, wherein said decay pattern maps the Reduction Level of some pixels relative to other pixels of described suspicious ROI of described suspicious ROI.
11. Computerized methods according to claim 1, wherein said generation comprises the described decay pattern of generation to map the decay in described suspicious ROI and peripheral region thereof.
12. Computerized methods according to claim 1, wherein said mark comprises by suspicious ROI described in user's labelling.
13. Computerized methods according to claim 1, wherein said US image is obtained by B-scan ultrasound therapy.
14. Computerized methods according to claim 1, wherein said measurement comprises by the following calculating at least one decay characteristics described:
Application notices that threshold value is to identify the multiple high attenuator region in described suspicious ROI;
Calculate multiple set, each set comprises multiple features of another subregion in described high attenuator region;
Calculate described Reduction Level variability measure by combining described multiple set.
15. Computerized methods according to claim 1, also comprise exporting and show that described classification is to the display of user.
16. Computerized methods according to claim 1, wherein said US image is three-dimensional (3D) US image and described suspicious ROI is suspicious 3D ROI; Wherein said generation comprises for each section generation of attenuation map in multiple sections of described suspicious 3D ROI; Wherein perform described measurement for each described section and perform described classification according to the combination of at least one decay characteristics each described of each described section.
17. 1 kinds of systems of classifying at least one suspicious regions of interest (ROI) in the ultrasonic attenuation image of the tissue of mapping patient, comprising:
Processor;
Input module, the US image of its tissue receiving;
ROI module, it identifies the suspicious regions of interest (ROI) in described US image;
Attenuation map module; It generates the decay pattern of described ROI; And
Sort module, it is classified to described suspicious ROI according at least one decay characteristics of at least one subregion in described suspicious ROI, and at least one decay characteristics described identifies according to the analysis of described decay pattern.
18. systems according to claim 17, wherein said tissue is the non-homogeneous tissue with non-homogeneous ultrasonic radiation absorption.
19. systems according to claim 17, also comprise the display for showing the described classification be associated with described US image.
20. 1 kinds of computer programs of classifying at least one suspicious regions of interest (ROI) in the ultrasonic attenuation image of the tissue to mapping patient, comprising:
Non-transient computer-readable recording medium;
First programmed instruction, for the US image of tissue receiving;
Second programmed instruction, for identifying the suspicious regions of interest (ROI) in described US image;
3rd programmed instruction, for generating the decay pattern of described suspicious ROI;
4th programmed instruction, at least one decay characteristics of at least one of at least one subregion and described suspicious ROI in ROI suspicious according to the analysis to measure of described decay pattern; And
5th programmed instruction, for classifying to described suspicious ROI according at least one decay characteristics described;
Wherein said first programmed instruction, described second programmed instruction, described 3rd programmed instruction, described 4th programmed instruction and described 5th programmed instruction are stored on described non-transient computer-readable recording medium.
CN201380027801.0A 2012-07-26 2013-06-21 Method and system for classifying suspicious area of interest in ultrasonic attenuation image CN104321020B (en)

Priority Applications (3)

Application Number Priority Date Filing Date Title
US13/558,372 2012-07-26
US13/558,372 US8781202B2 (en) 2012-07-26 2012-07-26 Tumor classification based on an analysis of a related ultrasonic attenuation map
PCT/CA2013/050486 WO2014015424A1 (en) 2012-07-26 2013-06-21 Tumor classification based on an analysis of a related ultrasonic attenuation map

Publications (2)

Publication Number Publication Date
CN104321020A true CN104321020A (en) 2015-01-28
CN104321020B CN104321020B (en) 2017-02-22

Family

ID=49994935

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201380027801.0A CN104321020B (en) 2012-07-26 2013-06-21 Method and system for classifying suspicious area of interest in ultrasonic attenuation image

Country Status (4)

Country Link
US (1) US8781202B2 (en)
CN (1) CN104321020B (en)
TW (1) TWI552013B (en)
WO (1) WO2014015424A1 (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20140093376A (en) * 2013-01-16 2014-07-28 삼성전자주식회사 Apparatus and method for estimating whether malignant tumor is in object by using medical image
WO2014150578A1 (en) * 2013-03-15 2014-09-25 Seno Medical Instruments, Inc. System and method for diagnostic vector classification support
KR20150080820A (en) * 2014-01-02 2015-07-10 삼성전자주식회사 Apparatus and Method for indicating region of interest
CN104143101A (en) * 2014-07-01 2014-11-12 华南理工大学 Method for automatically identifying breast tumor area based on ultrasound image
TWI574671B (en) * 2016-06-27 2017-03-21 太豪生醫股份有限公司 Analysis method for breast image and electronic apparatus thereof
US20190287234A1 (en) * 2016-12-06 2019-09-19 Siemens Energy, Inc. Weakly supervised anomaly detection and segmentation in images

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5627906A (en) * 1990-12-31 1997-05-06 International Business Machines Corporation Image processing system for producing an attenuation map of a scanned image
US5806521A (en) * 1996-03-26 1998-09-15 Sandia Corporation Composite ultrasound imaging apparatus and method
CN1989524A (en) * 2004-07-26 2007-06-27 皇家飞利浦电子股份有限公司 System and method for automated suspicious object boundary determination
CN101336831A (en) * 2008-08-13 2009-01-07 汕头超声仪器研究所 Rebuilding method of real-time three-dimensional medical ultrasonic image
EP2359745A1 (en) * 2010-02-12 2011-08-24 Helmholtz Zentrum München Deutsches Forschungszentrum für Gesundheit und Umwelt (GmbH) Method and device for multi-spectral photonic imaging

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6117080A (en) 1997-06-04 2000-09-12 Atl Ultrasound Ultrasonic imaging apparatus and method for breast cancer diagnosis with the use of volume rendering
US7058210B2 (en) * 2001-11-20 2006-06-06 General Electric Company Method and system for lung disease detection
US6878115B2 (en) 2002-03-28 2005-04-12 Ultrasound Detection Systems, Llc Three-dimensional ultrasound computed tomography imaging system
US7248725B2 (en) * 2004-01-07 2007-07-24 Ramot At Tel Avia University Ltd. Methods and apparatus for analyzing ultrasound images
US20100266179A1 (en) 2005-05-25 2010-10-21 Ramsay Thomas E System and method for texture visualization and image analysis to differentiate between malignant and benign lesions
US8834371B2 (en) 2005-03-31 2014-09-16 Kabushiki Kaisha Toshiba Ultrasound diagnostic apparatus and ultrasound image processing program
US8376947B2 (en) * 2008-03-26 2013-02-19 Bioquantetics, Inc. Application of image-based dynamic ultrasound spectrography (IDUS) in detection and localization of breast microcalcifcation
TWI399195B (en) * 2008-12-16 2013-06-21 Ind Tech Res Inst Apparatus and method for providing a dynamic 3d ultrasound image
CA2772777A1 (en) * 2009-09-18 2011-03-24 Kominox, Inc. Methods for treating brain tumors
TWI428781B (en) * 2010-01-25 2014-03-01 Amcad Biomed Corp Method for retrieving a tumor contour of an image processing system

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5627906A (en) * 1990-12-31 1997-05-06 International Business Machines Corporation Image processing system for producing an attenuation map of a scanned image
US5806521A (en) * 1996-03-26 1998-09-15 Sandia Corporation Composite ultrasound imaging apparatus and method
CN1989524A (en) * 2004-07-26 2007-06-27 皇家飞利浦电子股份有限公司 System and method for automated suspicious object boundary determination
CN101336831A (en) * 2008-08-13 2009-01-07 汕头超声仪器研究所 Rebuilding method of real-time three-dimensional medical ultrasonic image
EP2359745A1 (en) * 2010-02-12 2011-08-24 Helmholtz Zentrum München Deutsches Forschungszentrum für Gesundheit und Umwelt (GmbH) Method and device for multi-spectral photonic imaging

Also Published As

Publication number Publication date
TW201426379A (en) 2014-07-01
US20140029817A1 (en) 2014-01-30
TWI552013B (en) 2016-10-01
WO2014015424A1 (en) 2014-01-30
CN104321020B (en) 2017-02-22
US8781202B2 (en) 2014-07-15

Similar Documents

Publication Publication Date Title
Peng et al. Quantitative analysis of multiparametric prostate MR images: differentiation between prostate cancer and normal tissue and correlation with Gleason score—a computer-aided diagnosis development study
Kim et al. Statistical textural features for detection of microcalcifications in digitized mammograms
Panetta et al. Nonlinear unsharp masking for mammogram enhancement
Drukker et al. Computerized lesion detection on breast ultrasound
Armato et al. Lung cancer: performance of automated lung nodule detection applied to cancers missed in a CT screening program
Van Ginneken et al. Computer-aided diagnosis: how to move from the laboratory to the clinic
Kolb et al. Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations
US7313261B2 (en) User interface for computed tomography (CT) scan analysis
US8571292B2 (en) Breast tomosynthesis with display of highlighted suspected calcifications
Ganeshan et al. Quantifying tumour heterogeneity with CT
US7133546B2 (en) Digital medical image analysis
US7418123B2 (en) Automated method and system for computerized image analysis for prognosis
Chan et al. Computer-aided detection system for breast masses on digital tomosynthesis mammograms: preliminary experience
Giger et al. Breast image analysis for risk assessment, detection, diagnosis, and treatment of cancer
JP4949264B2 (en) System and method for automatically detecting and segmenting tumor boundaries in medical image data
US7840046B2 (en) System and method for detection of breast masses and calcifications using the tomosynthesis projection and reconstructed images
von Eyben et al. 68Ga-labeled prostate-specific membrane antigen ligand positron emission tomography/computed tomography for prostate cancer: a systematic review and meta-analysis
US6760468B1 (en) Method and system for the detection of lung nodule in radiological images using digital image processing and artificial neural network
US7646902B2 (en) Computerized detection of breast cancer on digital tomosynthesis mammograms
Banik et al. Detection of architectural distortion in prior mammograms
Keller et al. Estimation of breast percent density in raw and processed full field digital mammography images via adaptive fuzzy c‐means clustering and support vector machine segmentation
Chan et al. Computer‐aided detection of masses in digital tomosynthesis mammography: Comparison of three approaches
Shi et al. Characterization of mammographic masses based on level set segmentation with new image features and patient information
Zhou et al. Automatic multiscale enhancement and segmentation of pulmonary vessels in CT pulmonary angiography images for CAD applications
Reiser et al. Computerized mass detection for digital breast tomosynthesis directly from the projection images

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant